Richard Carrier and the Health Effects of Moderate Alcohol Consumption

Dr. Carrier on alcohol

Richard Carrier is a well-known skeptical historian and philosopher. He is the author of many great books, such as Sense and Goodness without God (where he outlines and defends philosophical naturalism) and a prolific writers of various articles on skepticism, origin of life, morality, the bible and history, Bayesian inference in historical research and social justice issues. I greatly enjoy reading Dr. Carrier, especially on topics such as skepticism, religion and morality and have devoured most of his articles, books and book-chapters on those topics. However, sometimes I find myself in less agreement with him. This post chronicles one such disagreement. Recently, Dr. Carrier wrote a blog post entitled My Favorite Scotch. In it, he discusses his taste in alcohol and muses over how it is that different people can have radically different tastes in alcohol. In the last few sections of the post, he debunks the myth that alcohol directly causes the death of neurons. He also claims the “science is clear” that moderate consumption of alcohol improves cognitive function, memory and brain function into old age.

As we will see, this is a clear case of cherry-picking and the science is actually not that clear at all regarding his claims about cognitive benefits. His source for those claims, a sociologist by the name of David Hanson, is an anti-psychiatry crank and promotes quack treatments for substance addiction. Furthermore, Dr. Carrier does not tell his readers that alcohol can indirectly cause the death of neurons via Wernicke–Korsakoff syndrome that is a result of thiamine deficiency due to chronic alcohol abuse. Dr. Carrier does not even explain the negative health consequences of moderate alcohol consumption, which includes elevated risk of cancer, injuries, violence, mortality in pancreatic inflammation and liver cirrhosis. When I take the time to explain this to Dr. Carrier (with references to the scientific literature), he responds with a condescending and overly rhetorical reply that did not discuss a single of my arguments. Despite that fact that I had backed up each major claim with a scientific reference, I was told that I had been “taken in my someone selling [that] narrative” and that I had “to bone up on [my] basic sociology and anthropology” as well as to “get out more”.

In his post, Richard Carrier made four core claims about the health effects of alcohol: alcohol does not kill neurons, moderate consumption of alcohol improves cognitive skill and memory, moderate alcohol consumption helps the brain function better during aging and only extreme abuse of alcohol can cause health problems. Carrier adds that “science is clear” with regards to the truth of the three latter claims.

Richard Carrier appeals to a false expert with overt ties to pseudoscience

In his attempt to justify some of these assertions (primarily the first three), Carrier references a person called David Hanson and post a couple of quotes, taking care to point out that Hanson has a Ph.D. However, Carrier fails to mention to his readers that the Ph.D. is not in a relevant scientific field. Rather, it is in sociology and Hanson is not even a professor at University of New York at Potsdam anymore. In terms of scientific qualifications in neurobiology, Hanson is just a guy on the Internet with a non-relevant degree.

It gets worse, because Hanson is a promoter and board member of an alternative treatment program for substance abuse called “The Saint Jude Retreat”. Despite stating that they are not a treatment, they claim a 62.5% success rate for long-term sobriety. In comparison, the treatment effect size for addiction using mainstream cognitive behavioral therapy is considered small to moderate. They also directly promote pseudoscience when they claim that alcoholism is not a disease (they call it a “choice”) and assert that there are no genetic risk factors or biological mechanisms for addiction.

In reality, substance addiction or abuse (whether alcohol or illicit drugs) often causes severe medical, social and economic functional impairment in the life of individuals and can thus legitimately be considered a disease (NIDA, 2010). There are also well-established neurological mechanisms that influence addiction and genetic risk factors have been reliably identified (NIDA, 2012; Sullivan et al, 2012).

Thus, Carrier’s arguments regarding the alleged health risks of alcohol consumption involves the appeal to a false expert. This is otherwise common among proponents of pseudoscience. Carrier made a substantial error in appealing to Hanson and it was unfortunate that Carrier did not properly check his sources before making that post. However, Richard Carrier is correct that alcohol does not directly kill brain cells. It is a well-known myth. Here is how a proper refutation of the myth should be done.

Alcohol does not directly kill neurons, but damages dendrites

It does not take that much effort to find reliable scientific sources debunking the myth that alcohol kills brain cells. I spent a couple of minutes on this question and could find at least two such sources (both freely and legally accessible online through various means). The first is a note in Scientific American Mind (2012):

Even in heavy drinkers, alcohol consumption doesn’t kill brain cells. It does, however, damage the ends of neurons, called dendrites, which makes it difficult for neurons to relay messages to one another.

The second is Lilienfeld, Lynn, Ruscio and Beyerstein (2010, p. 42), who in a similar fashion, although using a bit more popular phrasing, point out that:

Fiction: Alcohol kills brain cells.

Fact: Alcohol appears not to kill brain cells themselves, although it can damage neuronal “dendrites”, which are portals that bring messages into neurons.

Both sources state that alcohol does not kill brain cells directly, but they effect the communication between brain cells by damaging dendrites. No such caveat can be found in Carrier’s blog post. This is the first indication (and many more will come) that Carrier is being somewhat selective in his portrayal of the positive and negative health effects of alcohol consumption.

The effects of moderate consumption of alcohol on cognitive functioning

Dr. Carrier approvingly quoted Hanson claiming that alcohol improves cognitive functions and Dr. Carrier even stated that the “science is clear”. Is this really the case? Unfortunately for Dr. Carrier, the situation is more complex, equivocal and nuanced than that.

The literature reveals some studies that seem to suggest that moderate alcohol consumption improves cognitive functioning (Cervilla et al. 2000; Herbert et al. 1993 and others), whereas some other studies seem to indicate no relation between moderate alcohol consumption and cognitive function (Dent et al. 1997; Edelstein et al. 1998). However, many of the studies in both categories had insufficient control for confounders. Not only that, many of these studies had low statistical power. This affects the credibility of both kinds of studies. This means that results of the studies that found no relationship between alcohol and improved cognitive functioning might have been artifacts of low sample size. Had a higher sample size been used, they might have detected a positive relationship. It also means that some of the studies that show a positive relationship may be false positives. When the statistical power is low, studies that yield statistically significant relationships are likely to be overestimations (because only those studies have a high enough effect size to overcome the low sample size).

I tracked down two large and relatively recent studies in an attempt to shed light on this equivocal literature.

The first, Stampfer et al. (2005), looked at the association between moderate alcohol consumption and cognitive function in a set of 12000 female nurses and the change in cognitive function over time in around 11000 of those women. The multivariate analysis showed that nurses who drank 1.0-14.9 grams of alcohol per day (0-1 U. S. standard drinks) had a relative risk (compared with non-drinkers) of cognitive impairment of 0.77 (95% CI: 0.67-0.88) and the nurses who drank 15.0-30 grams of alcohol (~1-2 U. S. standard drinks) per day had a relative risk (compared with non-drinkers) of cognitive impairment of 0.78 (95% CI: 0.57-1.06). The latter estimate had very large 95% CI error bars. Thus, it would seem, that drinking some alcohol, but below 1 U. S. standard drink per day had a beneficial effect on cognitive functioning for these nurses during the study. However, the study had a couple of important limitations. The most important problem was that the study was only over two years, so it was too short to evaluate the relationship between drinking a moderate amount of alcohol and long-term cognitive decline. That means that this study enables us to say that moderate assumption of alcohol correlates with better cognitive functioning (controlled for a number of confounders) for this population (which only consisted of registrations nurses) during this short study period. These results are also consistent with moderate consumption of alcohol not causing cognitive impairment in the short-term.

The second study, Kesse-Guyot et al. (2012), looked at the effects of moderate alcohol consumption in a data set of over 3000 French adults after 13 years. This study has two distinct advances to other studies: it examined midlife exposure and the time between midlife exposure an neuropsychological evaluation was very long. Their results indicated that females who did not drink had a lower cognitive function than those women drinking 1-2 standard drinks per day (almost 2 tenths of a SD difference, fairly wide 95% CI from −3.29 to −0.2) and heavy male drinkers (more than three standard drinks per day) had better cognitive function than those that had 1-3 standard drinks per day (around 1 tenth of an SD difference, 95% CI 0.10-1.99). Two of the biggest limitation of this study was that there was no cognitive assessment at baseline and that the study population is unlikely to be representative of the general population.

So to sum up these two large studies: their mean results seem to indicate a certain cognitive benefit with moderate alcohol consumption, but the 95% confidence intervals were wide. This means that the suggested benefit is most likely small to moderate with a fairly large uncertainty. Some important limitations included the fact that the first study being too short and the second study having substantial issues with representativeness. That means that we cannot make rash extrapolations to the general population. Indeed, the second study (Kesse-Guyot et al. 2012) states explicitly that “There is substantial evidence from other health domains (e.g., cancer development) that even a single serving of alcohol might have a detrimental impact” and that “Given the known harmful effects of alcohol even in low doses regarding risk of cancer, the study does not provide a basis for modifying current public health messages.” Thus, it is by no means clear that the observed benefit overturns any medical cost/benefit analysis with respect to alcohol consumption.

In other words, we are not justified to claim, as Dr. Carrier does, that “the science is clear” with respect to the effects of moderate alcohol consumption of cognitive function. Rather, it is at best suggestive with some major limitations.

The effects of moderate alcohol consumption on memory and age-related dementia

Some of the claims that Dr. Carrier makes, such as moderate alcohol consumption helps the brain “function better during aging”, are very vague. Function better how? To investigate the exact nature of the claim, we turn to the post by Hanson that Carrier links, and we check the reference list. So what do we find?

The so called “scientific references” for all of the three claims (moderate consumption improved cognitive function, memory and “function better during aging”) consists of one press release, one conference presentation, two newspaper articles and four scientific papers. The the four scientific papers (Bates and Tracy, 1990; Anttila et al, 2004; Galanis et al, 2000 and Elias et al, 1999) are quite old and only three are related to old age, memory and/or dementia (the Bates and Tracy study is about cognitive impairment in young adults). However, serious problems exists with all of those studies.

One of them (Elias) perform as many as 42 comparisons divided across seven groups without any mention of a correction for multiple comparisons, thus suggesting that they may have capitalized on chance and only emphasized the minority of comparisons that yielded statistically significance.

Another (Anttila) defined alcohol consumption based on frequency and not amount, where infrequent was defined as “less than 1 drink per month” which is hardly the same as moderate consumption of alcohol according to conventional standards (1-2 standard drinks per day). Compared with infrequent drinkers, those that abstained completely had slightly less cognitive impairment when controlled for many confounders (rr = 0.91, 95% CIs: 0.42 to 1.83) and little over double the risk for dementia (rr = 2.15, 95% CIs: 1.01 to 4.59). However, the 95% CIs are unacceptably wide for both categories. For reduced cognitive impairments, likely estimates of the population parameter spans from greatly reduced risk (0.42) to almost double the risk (1.83). For risk of dementia, plausible values for the population parameter goes from no increased risk (1.01) to over 4.5 times the risk (4.59). Thus, any general conclusion is going to be highly problematic, even if we ignore the fact that the comparison was between 0 times per month and < 1 time per month and not between no alcohol consumption and moderate alcohol consumption.

The third study (Galanis) may have issues with representativeness. The study population was drawn from the surviving members of the Honolulu Heart Program/Honolulu-Asia aging Study (HHP/HAAS) cohort, containing Americans with Japanese ancestry. So the cohort itself is probably not representative for the overall population and surviving members of that cohort is likely to be even more unrepresentative (since it would then be a non-random selection from a non-random cohort).

So much for Hanson’s “references [to] the actual science”.

Let us look at some mode modern research. An article in Scientific American recommends in the title that people should drink two beers per day and that low to moderate alcohol consumption improves memory (Swaminathan, 2007). However, the actual study (Kalev-Zylinska, 2007) was done on transgenic rats who had unlimited and unregulated access to liquid with a moderate concentration of alcohol (2.5% w/v). Some differences were observed between wild-type rats getting 0% and 2.5% ethanol diet, but the sample size was only n = 6 per group and this, together with the fact that it was an animal study, cannot justify the claim that “the science is clear” with regards to the effects of moderate alcohol consumption on memory.

Another recent study is Weyerer et al (2011). They prospectively studied 3000 elderly Germans without dementia and investigated whether alcohol consumption was related to the incidence of dementia. The non abstinent group had a risk ratio for incidence of overall dementia of 0.71 (95% CIs spanning 0.53–0.96). These error bars are unacceptably wide, as plausible values for the effect ranges from fantastic (0.53) to negligible (0.96). They also did not correct for multiple comparisons. Out of the 18 comparisons made, only a minority was deemed statistically significant. Another major problem with this study was that about 50% of their random sample did not consent to be a part of their study, so their results may be influenced by severe selection bias. This means that this study has problems with representativeness.

One might argue that I am being too hard on these papers. However, let us recall that Dr. Carrier claim that the “science was clear” on these issues. I have shown that this is not the case.

Negative health effects of alcohol occur even without “extreme abuse”

In his blog post, Dr. Carrier categorically asserts that you “have to abuse alcohol to an extreme degree to suffer health problems from it”. Is this accurate? As it turns out, it is not. In fact, it is contradicted by mainstream science. Thus, the claim made by Dr. Carrier is not only false, but also misleading and dangerous.

A review paper published by Andréasson and Allebeck (2005) in Läkartidningen (one of the largest and most prestigious science journal in Sweden, dating back to 1904) gives us an overview of the positive and negative health effects of low to moderate consumption of alcohol (cognitive functions already discussed above):

  • Coronary artery disease: no obvious negative effects on heart and arteries, could be protective but interpretational problems obscure that conclusion. The benefit is on the same level as e. g. acetylsalicylic acid, but unlike this compound, no controlled clinical trials have been done on alcohol.
  • Alcohol-related accidents: alcohol is globally responsible for around 30% of accidental deaths in the major accident categories and a substantial portion of these are due to low to moderate alcohol consumption.
  • Cancer: alcohol is strongly associated to many different cancer forms, including cancer in the oral cavity, pharynx, esophagus and larynx. It is also less strongly related to cancer in the stomach, colon and rectum as well as liver and breast cancer. These associations exist even for moderate alcohol consumption (~2 U. S. standard drinks per day) and the relative risk is between 1.2-2 for that consumption. For this section, Andréasson and Allebeck reference Bagnardi et al. (2001).
  • Diabetes: a moderate alcohol consumption reduces the risk of type-2 diabetes with around 20-30% compared with those that drink very little or not at all.
  • Cirrhosis of the liver: a moderate alcohol consumption can triple the risk of cirrhosis of the liver. See for instance, Corrao et al. (1998) and later studies such as Rehm et al. (2010) also found an increase risk with moderate consumption, although the risk increase is exponential the higher the consumption becomes.
  • Mortality in pancreatic inflammation: alcohol is the most common reason for inflammation of the pancreas and it has been linked to moderate alcohol intake as well.
  • Pregnancy: drinking alcohol while pregnant has been shown to be associated with problems with learning. These effects have been found to be dose-dependent and an effect exists even at moderate alcohol consumption. For this section, Andréasson and Allebeck reference Streissguth et al. (1994).

In addition, the authors reference Norström and Skog (2001), which show, among other things, that moderate alcohol consumption is linked to accidents, cirrhosis, homicide and all-cause mortality.

They also point out that the beneficial effects of moderate alcohol consumption are quantitatively small and that studies that appear to show health benefits have many methodological problems such as (1) comparisons are clouded by the fact that some people who do not drink do so because of psychosocial issues or disease and (2) some studies put people who do not drink at all together with people who rarely drink and it is known that this latter group has a higher burden and worse prognosis than the former (this is attributed that some older individuals who drink a lot become sick and decrease their alcohol consumption substantially).

For a more authoritative review of the health consequences of moderate alcohol consumption, please see State of the Science Report on the Effects of Moderate Drinking (NIAAA, 2003), which is more detailed and has a lot more references to the scientific literature.

Thus, Dr. Carrier is wrong when he claims that negative health effects of alcohol can only occur with “extreme abuse” is falsified by the data.

Case study: a short interaction with Dr. Carrier

Self-deception occurs when a person rejects or dismiss evidence and arguments that run counter to their beliefs and positions. Together with other well-known cognitive biases and logical fallacies, such as confirmation bias and cherry-picking, self-deception has the potential to become incredibly powerful. Besides the context of pseudoscience, self-deception features prominently in the use of recreational drugs. A recreational drug is a substance that is consumed partly because of its psychoactive properties and effects and this category includes not only illicit drugs, but also alcohol, tobacco and coffee (WHO, 2013).

It is my contention that it can sometimes be very difficult to have a rational discussion about a recreational drugs and their health effects with a user because some users. My short interaction with Dr. Carrier in the comment section of his post is an interesting case study regarding that issue to which we now turn our attention.

A short exchange between myself and Dr. Carrier

…this was not quite the response I was expecting from Dr. Carrier.

In my comment, acknowledged that Dr. Carrier was right on some issues and wrote a relevant comment on the topic of health effects of moderate alcohol consumption. I pointed out that he had repeatedly treated the issue selectively. I even references the scientific literature and the U. S. Centers for Disease Control and Prevention to back up my claims about the negative effects of alcohol on society and the negative effects of moderate alcohol consumption on the individual. I expected a thoughtful and relevant rebuttal that discussed the issues I brought up, an admission that he had been somewhat selective in his treatment of the health effects of moderate alcohol consumption and perhaps even a more detailed discussion of moderate alcohol consumption on his part.

Instead, I got a vacuous and overly rhetorical reply that did not discuss a single of my arguments. Despite that fact that I had backed up each major claim with a scientific reference, I was told that I had been “taken in my someone selling [that] narrative” and that I had “to bone up on [my] basic sociology and anthropology”. Finally, I was told to “get out more”.

At that point, I decided that further interaction with Dr. Carrier on this matter would probably not be productive. Instead, let us examine each segment of Dr. Carrier’s reply in detail.

Have I been taken in by someone selling a narrative?

I find this objection difficult to understand. It does not address any of the arguments or evidence I presented. It does not present any arguments or evidence on its own. Instead, it is the assertion that a “narrative” (presumably not based on evidence) related to the health effects of moderate alcohol consumption exists and that I have somehow “been taken in”.

Such assertions seem vaguely conspiratorial when they are used without evidence against a position that is backed by evidence. That, however, seems very out-of-character for Dr. Carrier. After reading some more of the writing of David Hanson (the person Dr. Carrier favorably cited earlier), I found a post called Repeal of National Prohibition, where Hanson stated that:

Their [“Modern prohibitionists” – Emil’s note] tactic is to establish cultural rather than strictly legal prohibition by making alcohol beverages less socially acceptable and marginalizing those who drink, no matter how moderately. Like the hatchet-wielding Carrie Nation and other prohibitionists who preceded them, modern prohibitionists (or neo-prohibitionists) don’t distinguish between the use and the abuse of alcohol.”

Thus, it seems that Dr. Carrier has gotten his ideas about being “taken in” by someone promoting that “narrative” from Hanson (who seem stuck in the 1920s). As we saw above, Hanson is some sort of anti-psychiatry crank who cannot be considered scientifically credible. Ironically, it seems that it is Dr. Carrier (who relies on a quack), not me (who rely on peer-reviewed science and the CDC), that has been taken in by someone selling a narrative. This is also apparent from Dr. Carrier’s assumption that I am somehow anti-alcohol (claiming that I cannot tell the difference between irresponsible drinking and mere consumption, telling me to get out more etc.), thus placing me, without justification, in the category of so-called “modern prohibitionists”.

Wernicke–Korsakoff syndrome was an example to point out Dr. Carrier’s cherry-picking

The main point of my comment was to point out that Dr. Carrier’s treatment of the effect of alcohol had been very selective. He pointed out that the claim that alcohol directly causes the death of neurons was false, but he failed to point out that long-term chronic abuse of alcohol lead to the death of neurons with Wernicke–Korsakoff syndrome caused by thiamine deficiency. He did not even explain to his readers that alcohol can cause damage to dendrites and affect intercellular communication between neurons.

Dr. Carrier seems to have interpreted my claim as saying that moderate consumption of alcohol cause Wernicke–Korsakoff syndrome. However, nowhere in my comment do I claim this. My comment about Wernicke–Korsakoff syndrome was an effort to counteract Dr. Carrier’s selective treatment of the negative health effects of alcohol.

Even moderate consumption of alcohol elevates cancer risk 20-100% for nine cancer types

In his “response”, Dr. Carrier implies that more than moderate alcohol consumption is needed to cause “a significant elevation of cancer risk”. If Dr. Carrier means a statistically significant elevation, then his position is falsified by the research referenced in this post. The risk of cancer in oral cavity, pharynx, esophagus, larynx, stomach, colon, rectum, liver and breast cancer is statistically significantly elevated in individuals who engage in a moderate consumption of alcohol (Bagnardi et al, 2001).

If Dr. Carrier means a practically significant elevation, then we have to look at the relative risks reported. For the 25 g / day category (around 2 standard drinks; moderate consumption) in the Bagnardi paper, the relative risks for different forms of cancer are around 1.2 (20% increased risk) to around 2 (100% increased risk i. e. twice the relative risk) compared with non-drinkers. That means that the population of moderate drinkers have up to twice the relative risk of some cancer types than non-drinkers.

Now, it is possible to argue that the baseline risk of these cancers are very low, so even a doubling of the risk is going to be practically negligible. That might be a reasonable argument if the elevated risk was only for a single cancer type, but the risk elevation occurs for at least nine different types of cancer. So the accumulated increase in cancer incidence for all of those cancer forms may not be defeated by that objection.

Another argument is that it is possible to accept this increased risk because a person feels that a moderate alcohol consumption contributes more to the quality of life for that person than the increased risk reduce it. That is an acceptable argument, but it is important to think hard about exactly what aspects of moderate alcohol consumption increase quality of life. Is it more related to the psychoactive effects of alcohol or more related to the social aspects of alcohol consumption with friends? If it is the later, there may be other ways to accomplish that without actually consuming alcohol.

Alcohol-related deaths not exclusively related to irresponsible drinking

Dr. Carrier implicitly claim that it is irresponsible drinking that is the cause of alcohol-related deaths and not the moderate consumption of alcohol. However, as we saw in Andréasson and Allebeck (2005), a substantial proportion of the alcohol-related accident are due to low to moderate consumption. Thus, those injuries and deaths was not due to irresponsible drinking. It might be claimed that it was by definition irresponsible drinking since it led to injuries and deaths, but that is irrelevant. The relevant part here is that a lot of these injuries and deaths are attributable to moderate consumption of alcohol, whatever you chose to call it. In fact, Vinson et al (2003) found that as little as 1 to 2 drinks doubles the injury risk.

The report from the National Institute of Alcohol Abuse and Alcoholism cited earlier (NIAAA, 2003) has this to say about the effects of moderate drinking and injuries/accidents.


Studies on the role of alcohol in injury from falls and violence/abuse frequently do not distinguish between moderate and excessive drinking (e.g., Cunningham et al., 2003; Humphrey et al., 2003; Vinson et al., 2003; Wells & Graham, 2003; Zautcke et al, 2002). However, many “moderate drinkers” have episodes of high-risk drinking, including heavy episodic drinking and acute intoxication leading to injuries and violence (Gutjahr et al., 2001). Additionally, studies of the acute effects of alcohol show that even moderate-dose consumption compromises brain performance in terms of error detection, processing speed, and response time (Ridderinkhof et al., 2002), impairments that may be particularly important in terms of driving-related risk. Several reports (Deery & Love, 1996; Hingson et al., 1999; Midanik et al. 1996) have indicated that low levels of drinking (e.g., 1 or fewer per day) and BACs below the legal limit of 0.08% (e.g., 0.05%) increase risk of driving-related accidents.

Dr. Carrier cannot arbitrarily exclude one category of negative health effects because you think it is irresponsible. In the end, the amount was still within moderate consumption or less, thus making it justifiable to include these injuries and deaths into the category of negative health consequences of moderate consumption.

The effect of moderate alcohol consumption of pregnancy

In my comment, I pointed out that even moderate alcohol consumption increases the risk of damage to the fetus. Dr. Carrier responds by saying that this is only a problem if you have a fetus in you. This is of course an irrelevant objection. When discussing the negative health effects of moderate alcohol consumption, it is legitimate to take into account all categories of effects, including those that affect fetuses. Thus, Dr. Carrier dismisses and arbitrarily excludes one category of negative health effects without any justification.

Alcohol consumption is related to violence

Dr. Carrier apparently finds it amusing when people point out that scientific research shows that moderate alcohol consumption is linked to violence. He dismisses this and encourages me to read up on sociology and anthropology (insinuating that people who disagree with him must be ignorant of the relevant science). This suggests that he does not accept that alcohol facilitates violence, which is strange since he conceded that if you drink enough to become drunk, that “produces a state of global cognitive impairment”. This should reasonably involve reduction in self-control and a decreased ability for rational conflict resolution.

Nevertheless, let’s follow Dr. Carrier’s advice: here is what the 6th edition of Vold’s Theoretical Criminology (Bernard, Snipes and Gerould, 2010, pp. 52-53), one of the worlds leading textbooks on criminology, has to say about alcohol and violence:

Alcohol is known to increase aggressive behavior temporarily in lower doses (when people get nasty) and temporarily decrease aggressive behavior in higher doses (when people pass out). Many people believe that the increased aggressiveness at lower doses occurs because alcohol tends to release people from their inhibitions, but there is little evidence for this belief. An alternative explanation is that alcohol increases the production of the endocrine system, especially testosterone, but again, there is little evidence for this explanation.

Serotonin is a possible mediating factor between alcohol and violence. Because alcohol is known to decrease serotonin, and low serotonin is known to increase aggression, it seems logical to infer that low serotonin is one reason why alcohol increases aggression. Research has found physiological evidence both for and against this mediating role for serotonin in the relationship between alcohol and aggression, but “a relatively much larger body of evidence exists in support of such a role”. Other researchers believe that there may be a genetic basis for the relationship between alcohol and violence, but there has been no confirmation of this belief to date. So while the relationship between alcohol and violence is probably the strongest of any drug, especially for males, the reason for this relationship remains unclear.

In fact, the textbook go so far as to explicitly list the ingestion of alcohol as one of the nine major factors that increases a person’s likelihood of committing criminal behavior (p. 353). The NIAAA report cited above paints a similar picture with respect to moderate alcohol consumption and injuries due to violence.

…get out more?

The final part of Dr. Carrier’s reply to me was that I should “get out more”. Like the first part about how I have supposedly been taken in by a narrative, this complaint is difficult to understand from the perspective of a rational conversation about the negative health effects of alcohol. Not only does Carrier assume that I do not “get out” sufficiently (according to some standard), but it is not that clear why the frequency with which I do “get out” (presumably Dr. Carrier means the frequency of visits to establishments that sell alcohol) is in any way relevant to the reasonableness of my arguments or the accuracy of the research that I cite.

I have actually gotten this objection quite frequently when I discuss the positive and negative health effects of alcohol. Although I do not know Dr. Carrier’s intention, the usual story is that if I were to just “get out more” (presumably consume more alcohol in social situations) then I would realize how wrong I am on this issue. From my perspective, this is not much different from the claim that I would be less critical of homeopathy if I just started to consume it moderately.

Thus, Dr. Carrier has fallen to the stereotype that people who accept that the medical cost/benefit analysis does not favor (and possibly slightly disfavor) the consumption of a moderate amount of alcohol must be dull, lonely and socially isolated in the sense of not “getting out” enough. However, it is important to understand that it is entirely possible to have a vibrant social life and visit such establishments frequently without consuming alcohol (choosing instead to drink other non-alcoholic beverages) or merely consuming a low amount of alcohol. It is ironic that Dr. Carrier responded like this just days after going on FtBCon talking about the problems with the stereotyping of women in the coast guard and recounting how his stereotypes were dispelled.


While Dr. Carrier is correct that alcohol does not kill neurons directly and that moderate consumption of alcohol may give a small benefit in terms of cognitive function, he cherry-picked the scientific data on the health consequences of moderate consumption.

A fuller view of the health consequences of moderate alcohol consumption reveal a reduction in the risk for type 2 diabetes, a small potentially protective benefit against coronary artery diseases, an increase risk of at least nine cancer types, increase in risk of accidents as well as violence, an increased risk of liver cirrhosis, increased mortality in pancreatic inflammation and damage to the fetus. The benefit for cognitive function is small and those studies have important limitations, such as issues with representativeness. Thus, Dr. Carrier is wrong when he claims that the “science is clear” on the effects of moderate alcohol consumption on cognitive function.

Dr. Carrier is also wrong in his claim that negative health effects require extreme abuse and that moderate consumption of alcohol is unrelated to violence.

In his cherry-picking efforts, Dr. Carrier relies exclusively on David Hanson, a retired sociologist who sits on the board of The Saint Jude Retreat, an organization providing people with crank treatments for substance addiction, claim wild success rates uncorroborated by any published research and promoting various anti-psychiatry claims, such as calling alcoholism “a choice” instead of a disease and the rejection of the existence of biological risk factors and neurological mechanisms for addiction.

The studies provided by Hanson had serious limitations, including non-representativeness, large error bars, failure to correct for multiple comparisons and radically non-standard definitions of moderate consumption. Even if Dr. Carrier was accurate in his claim that moderate consumption provides these small benefits, most papers and reports reviewed in this article does not think it justifies a change in the medical guidelines for alcohol consumption.

In the end, it was highly inappropriate for Dr. Carrier to rely on Hanson for facts about the health effects of moderate alcohol consumption and his refusal to engage the actual arguments I presented was unfortunate.

I have no problem with people who want to engage in a moderate consumption of alcohol and that should certainly be allowed. What is not alright is the selective presentation of certain alleged positive health effects in an attempt to defend such a consumption level on medical grounds.


Andréasson S, Allebeck P. (2005). Alcohol as medication is no good. More risks than benefits according to a survey of current knowledge. Lakartidningen. 102(9):632-7.

Anttila, Tiia, Helkala, Eeva-Liisa, Viitanen, Matti, Kåreholt, Ingemar, Fratiglioni, Laura, Winblad, Bengt, . . . Kivipelto, Miia. (2004). Alcohol drinking in middle age and subsequent risk of mild cognitive impairment and dementia in old age: a prospective population based study. BMJ, 329(7465), 539.

Bagnardi, V., Blangiardo, M., Vecchia, C. La, & Corrao, G. (2001). A meta-analysis of alcohol drinking and cancer risk. Br J Cancer, 85(11), 1700-1705.

Bates M.E., Tracy J.I. (1990) Cognitive functioning in young “social drinkers”: is there impairment to detect? J Abnorm Psychol. Aug;99(3):242-9.

Bernard, Thomas J., Snipes, Jeffrey B., & Gerould, Alexander L. (2010). Vold’s Theoretical Criminology (International Sixth ed.). New York: Oxford University Press.

Cervilla J.A., Prince M., Mann A. (2000). Smoking, drinking, and incident cognitive impairment: a cohort community based study included in the Gospel Oak project. J Neurol Neurosurg Psychiatry. 68:622-626

Corrao, G. Bagnardi, V. Zambon, A., Torchio, P. (1998). Meta-analysis of alcohol intake in relation to risk of liver cirrhosis. Alcohol and Alcoholism, 33(4), 381-392.

Dent O.F., Sulway M.R., Broe G.A., et al. (1997). Alcohol consumption and cognitive performance in a random sample of Australian soldiers who served in the Second World War. BMJ. 314:1655-1657

Edelstein S.L., Kritz-Silverstein D., Barrett-Connor E. (1998). Prospective association of smoking and alcohol use with cognitive function in an elderly cohort. J Womens Health. 7:1271-1281

Elias, Penelope K., Elias, Merrill F., D’Agostino, Ralph B., Silbershatz, Halit, & Wolf, Philip A. (1999). Alcohol Consumption and Cognitive Performance in the Framingham Heart Study. American Journal of Epidemiology, 150(6), 580-589.

Galanis D.J., Joseph C., Masaki K.H., Petrovitch H., Ross G.W., White L. (2000). A longitudinal study of drinking and cognitive performance in elderly Japanese American men: the Honolulu-Asia Aging Study. Am J Public Health. 90(8):1254-9.

Herbert L.E., Scherr P.A., Beckett L.A., et al. (1993). Relation of smoking and low-to-moderate alcohol consumption to change in cognitive function: a longitudinal study in a defined community of older persons. Am J Epidemiol 137:881-891

Kalev-Zylinska, Maggie L., & During, Matthew J. (2007). Paradoxical Facilitatory Effect of Low-Dose Alcohol Consumption on Memory Mediated by NMDA Receptors. The Journal of Neuroscience, 27(39), 10456-10467. doi: 10.1523/jneurosci.2789-07.2007

Kesse-Guyot, Emmanuelle, Andreeva, Valentina A., Jeandel, Claude, Ferry, Monique, Touvier, Mathilde, Hercberg, Serge, & Galan, Pilar. (2012). Alcohol Consumption in Midlife and Cognitive Performance Assessed 13 Years Later in the SU.VI.MAX 2 Cohort. PLoS ONE, 7(12), e52311. doi: 10.1371/journal.pone.0052311

Lilienfeld, Scott O., Lynn, Steven Jay, Ruscio, John, & Beyerstein, Barry L. (2011). 50 Great Myths of Popular Psychology: Shattering Widespread Misconceptions about Human Behavior. West Sussex: Wiley-Blackwell.

NIAAA. (2003). State of the Science Report on the effects of moderate drinking. Accessed: 2013-07-25.

NIDA. (2010). Is drug addiction a mental illness? . Accessed: 2013-07-26.

NIDA. (2012). DrugFacts: Understanding Drug Abuse and Addiction. Accessed: 2013-07-26.

Norström T., Skog O.J. (2001). Alcohol and mortality: methodological and analytical issues in aggregate analyses. Addiction. 96 Suppl 1:S5-17.

Rehm, Jürgen, Taylor, Benjamin, Mohapatra, Satya, Irving, Hyacinth, Baliunas, Dolly, Patra, Jayadeep, & Roerecke, Michael. (2010). Alcohol as a risk factor for liver cirrhosis: A systematic review and meta-analysis. Drug and Alcohol Review, 29(4), 437-445. doi: 10.1111/j.1465-3362.2009.00153.x

Scientific American Mind. (2012). Drinking alcohol in excess kills brain cells. Scientific American. 23. 10.

Stampfer, Meir J., Kang, Jae Hee, Chen, Jennifer, Cherry, Rebecca, & Grodstein, Francine. (2005). Effects of Moderate Alcohol Consumption on Cognitive Function in Women. New England Journal of Medicine, 352(3), 245-253. doi: doi:10.1056/NEJMoa041152

Streissguth A.P., Barr H.M., Olson H.C., Sampson P.D., Bookstein F.L., Burgess D.M. (1994). Drinking during pregnancy decreases word attack and arithmetic scores on standardized tests: adolescent data from a population-based prospective study. Alcohol Clin Exp Res 18:248-54.

Sullivan, Patrick F., Daly, Mark J., & O’Donovan, Michael. (2012). Genetic architectures of psychiatric disorders: the emerging picture and its implications. Nat Rev Genet, 13(8), 537-551.

Swaminathan, N. (2007). Don’t Forget: Drink a Beer—Or Two—Daily!. Accessed: 2013-07-26.

Vinson D.C., Maclure M., Reidinger C., Smith G.S. (2003). A population-based case-crossover and case-control study of alcohol and the risk of injury. J Stud Alcohol. 64(3):358-66.

Weyerer, Siegfried, Schäufele, Martina, Wiese, Birgitt, Maier, Wolfgang, Tebarth, Franziska, van den Bussche, Hendrik, . . . group, for the German AgeCoDe Study. (2011). Current alcohol consumption and its relationship to incident dementia: results from a 3-year follow-up study among primary care attenders aged 75 years and older. Age and Ageing, 40(4), 456-463.

WHO. (2013). Lexicon of alcohol and drug terms published by the World Health Organization. Accessed: 2013-07-26.


Debunker of pseudoscience.

10 thoughts on “Richard Carrier and the Health Effects of Moderate Alcohol Consumption

  • “In reality, substance addiction or abuse (whether alcohol or illicit drugs) often causes severe medical, social and economic functional impairment in the life of individuals and can thus legitimately be considered a disease (NIDA, 2010).” That’s a non sequitur. A car accident that paralyses you from the waist down is not a disease. Long-term unemployment is not a disease. Plus, the citation, http://www.drugabuse.gov/publications/research-reports/comorbidity-addiction-other-mental-illnesses/drug-addiction-mental-illness is a FAQ-type answer that never mentions the word disease. Rather it claims drug addiction is a mental illness and compares it to a disorder.

    • All mental illnesses are diseases. Phrases like “illness”, “disorder” and “disease” are often used interchangeably. Thus, my citation supports the contention that substance addiction is a disease. You can also check out DSM-V and ICD-10, who both list substance addiction.

      My conception of a disease has no problem including partial paralysis, so that alleged counterexample cannot be accepted. Unlike long term unemployment, substance addiction (and partial paralysis) comes with a definitive pathology. This distinction invalidates your second proposed counterexample.

      At any rate, a complete definition of the disease concept do not need to be proposed to be justified in stating that a certain condition qualifies as a disease. We may not be able to tell for sure where the head ends and the neck begins, but we can surely state that the eyes are part of the head and not the neck. Thus, we can reliably say that substance addiction is a disease even though we have not quite nailed down a complete definition of the disease concept that include all diseases and exclude all non-diseases.

    • When talking about mental illness or other brain conditions like autism, disorder and disease are not synonyms. In fact, “disorder” is used to exclude the common conceptions of disease, such as “cure.”

      An event like a car accident is not a disease. The condition it causes and the prognosis may be the same as someone paralysed from a disease, but that doesn’t make it a disease.

    • In fact, “disorder” is used to exclude the common conceptions of disease, such as “cure.”

      That is not accurate.

      There are many conditions called diseases that do not have a cure, such as genital herpes, Parkinson’s disease and multiple sclerosis.

      There are also many conditioned labeled as a disorders that can be cured, such as specific phobias (an anxeity disorder) or X-linked severe combined immunodeficiency (an immunodeficiency disorder curable by gene therapy).

      An event like a car accident is not a disease. The condition it causes and the prognosis may be the same as someone paralysed from a disease, but that doesn’t make it a disease.

      No one is claiming that a car accident is a disease. You are performing a straw man fallacy. What I claimed was that I had no problem considering partial paralysis (regardless of the precise cause) to be a disease condition.

      I can also present further evidence that terms like “disease”, “disorder”, “illness” can be used interchangeably. For instance, the National Institute of Mental Health has this to say:

      disease: A synonym for illness. See illness.

      disorder: An abnormality in mental or physical health. In this module, disorder is used as a synonym for illness.


      illness: A problem in which some part or parts of the body do not function normally, in a way that interferes with a person’s life. For the purpose of this module, other terms considered to be synonyms for illness include disease, disorder, condition, and syndrome.

      This means that your position finds itself in a lot of trouble. To be considered credible, you must show why we should accept your statements rather than the statements of mainstream medicine. You shoulder a heavy burden of evidence.

      Not to mention that you have yet to provide any evidence whatsoever for your initial claim that substance addiction / abuse is not a disease. If your claim differs from that of the anti-psychiatry proponent Hanson (i.e. if your objection constitutes a quibble about semantics rather than a denial of mental illness), then such a position is still falsified by the National Institute of Mental Health who treats the terms synonymously.

  • It is also a heavy burden to incorrectly cite a source, as I showed that you did in my first comment. Once you’ve corrected this, I’ll address your other points.

    • The source is correctly cited. The initial reference I cited states that substance addiction is a mental illness. That is enough to show that substance addiction is a disease. You required additional evidence and I showed that the NIMH considers illness, disorder and disease as synonymous. I also refuted your claim that “disorder” specifies conditions without a cure by providing multiple counterexamples and exposed your straw man that I considered a car accident to be a disease.

      At this point, you have not addressed any of my arguments, but merely repeated your initial assertion. I suggest you review the comment guidelines on this blog before making your next comment on this blog. Pay special attention to point 7.

  • Pingback: David Brooks: Not Beyond Neuroscience Denial | Debunking Denialism

  • Have you forwarded this to Dr. Carrier? I’d be interested in his response.

    • Based on the content and attitude of his reply to my comment, I decided that further interaction with Dr. Carrier on this matter is unlikely to be productive.

  • I suspect that the claimed benefits of alcohol are only applicable to situations where something else is wrong (alcohol as an alternative to some other, worse health problem), while if all other factors are done right, then alcohol is always harmful.

    A good, well known example, that should be familiar to Richard Carrier as he is an ancient historian, is that in the antiquity water supply was not clean but contaminated with germs. Thus drinking alcohol was “beneficial” relative to drinking unclean water. Today in developed countries water is clean, so drinking alcohol is harmful compared to drinking water.

    So a test scenario could be: there is a broken diet where most of the energy comes from high-glyceminc-intex refined carbohydrates, Thus it causes excessive sugar ups and lows increasing chances for diabetes, killing of neurons due to high glucose levels, and low cognitive abilities during sugar lows. Take this diet, replace some amount of calories from carbohydrates with a similar amount of calories from regular consumption of alcohol. My bet would be that both cognitive, and diabetes-related results would improve – but that both diets would loose to a diet in which the secondary sources of calories for brain and body is not alcohol, but ketones delivered by the consumption of MCT oil (that is compared to doing the right thing).

    If so, then since our modern diets contain too much refined carbohydrates, and MCT oil is not very common among normal people (not sportsmen/bodybuilders/health enthusiasts/biohackers etc.) it would mean that observational studies probably have wrong control groups: normal people, that is people who do something wrong. Thus a testers and controls for testing effects of alcohol consumption should contain people who practice caring about health and do everything right. I suspect that if you give alcohol to such people then their health would degrade, rather than improve.

Comments are closed.

%d bloggers like this:

Hate email lists? Follow on Facebook and Twitter instead.